17 Full-Field 3D Experimental Modal Analysis from Dynamic Point Clouds Measured Using a Time-of-Flight Imager 165 Fig. 17.2 Very high-resolution 3D mode shapes (number of points equal to the number of virtual sensors), modal coordinates, and their power spectral densities (PSD) estimated using the imager-based blind identification approach 17.3 Analysis Laboratory tests were carried out with a stainless steel plate measuring 65 cm long by 30.5 cm wide. Data acquisition was performed using a Microsoft Azure Kinect camera featuring a ToF sensor. For comparison purposes, three laser displacement sensors were used along with an oscilloscope to estimate displacements at different locations of the plate. The excitation source is a modal hammer, used to impact the plate with a single hit. From the proposed technique, 90.186 virtual sensors were defined throughout the point cloud data (on average, each dynamic cloud has 110.000 points). From the estimated motion time series, the first three individual modes estimated are as follows: first bending mode (at 1.30 Hz), first torsional mode (at 3.91 Hz), and the second bending mode (at 7.56 Hz). In comparison, from the laser data, the same modes were estimated with the following frequencies: 1.35, 4.10, and 7.45 Hz. The recovered mode shapes and modal coordinates are visualized in Fig. 17.2. 17.4 Conclusions This paper presented a novel formulation for video-based modal analysis and identification using dynamic point cloud data. The BSS paradigm is utilized for unsupervised and output-only parameter estimation. As dynamic point clouds are inherently 3D data sets, the end result is the estimation of 3D high-resolution full-field modal parameters. Here, polynomial functions are used for fitting the data due to its simplicity and broad use in other related applications, such as finite element analysis. Alternatively, other models could be employed; radial basis functions are a suitable alternative to fit more complicated shapes to data. To the authors’ best knowledge, the processing of dynamic point clouds for modal analysis has never been attempted before, which endorses the unique contributions here stated. Acknowledgments This study was financed in part by the CAPES - Brazil - Finance codes 88882.445119/2018-01 and 88881.190499/2018-01. This work was supported by the US Department of Energy through the Los Alamos National Laboratory. Los Alamos National Laboratory is operated by Triad National Security, LLC, for the National Nuclear Security Administration of US Department of Energy (Contract No. 89233218CNA000001). This work was partially funded by the US Department of Energy Microreactor Project. References 1. Mistretta, F., Sanna, G., Stochino, F., Vacca, G.: Structure from motion point clouds for structural monitoring. Remote Sens. 11, 1940 (2019) 2. Jo, H.C., Sohn, H.-G., Lim, Y.M.: A lidar point cloud data-based method for evaluating strain on a curved steel plate subjected to lateral pressure. Sensors. 20, 721 (2020)
RkJQdWJsaXNoZXIy MTMzNzEzMQ==